These are the tactics that make the biggest difference on an ordinary day, assuming you already know what the rank means and how the bands work. If you do not, read that first — everything here is built on it. And if today’s board is still empty, play today’s puzzle alongside this page; the tips land better applied than remembered.
Opening words that earn their place
Your first three guesses are not attempts to win. They are measurements, and the words that measure best are the ones that sit far apart from each other in meaning-space.
1. Open with three words from unrelated domains. A good default trio: one concrete physical object, one abstract human concept, one natural phenomenon. For example keyboard, memory, storm. Whichever comes back warmest tells you which third of the space to search, and the other two are cheap because they cost nothing but a line on the board.
2. Use words with dense neighbourhoods. Common, well-connected words return more useful ranks than rare ones. Water, music, money, body, time, house, animal, war, light and child each sit at the centre of a large cluster, so a mid-range rank from one of them is a real signal rather than noise.
3. Do not reuse yesterday’s answer as an opener. It is tempting and almost always wasted — consecutive answers are chosen independently, so yesterday’s word carries no information about today’s.
4. Keep the same opening set for a couple of weeks. Consistency is worth more than optimality. Once you have seen storm come back as 900, 4,000 and 18,000 on different days, you develop a feel for what each number implies. A new opener every day gives you nothing to compare against.
Reading a rank properly
5. Treat anything past 1,500 as a single value: “no”. The difference between 4,000 and 11,000 is not a gradient you can follow. Both mean wrong region. Players who try to optimise inside the red band routinely spend thirty guesses going nowhere.
6. Treat a jump into the top 300 as a mode change. The moment something goes green, stop generating new ideas and start working the one you have. The exploration phase is over and its habits become expensive.
7. Compare, don’t just read. A rank is only meaningful next to your other ranks. Night at 55 tells you little on its own; night at 55 while sky is at 68 and planet is at 104 tells you the answer is specific rather than categorical, which is a real deduction.
8. Watch for a plateau. If four consecutive guesses all land between 200 and 400, you have found the category but not the axis inside it. Change the kind of word you are guessing — from noun to verb, from object to property — rather than trying more members of the same set.
Tactics for the green band
9. Guess the antonym of your best word. Opposites appear in nearly identical sentences, so the model places them close together. If cold is at 30, hot is probably also near 30 — and one of them may be adjacent to the answer in a way the other is not. It is one guess for a large amount of information.
10. Climb and descend the category ladder. From oak, try tree (up a level) and acorn (down a level). Abstraction level is one of the strongest axes in an embedding, and moving along it deliberately is far more efficient than guessing sideways among siblings.
11. Guess the verb. Most players guess nouns almost exclusively. When an answer is an activity, a state or an artefact defined by its purpose, the verb attached to it often ranks dramatically better than any noun. A lullaby round cracks open on soothe, not on song.
12. Guess the thing it is made of, and the thing it does. For physical answers, materials and functions are two reliably high-ranking directions: glass and reflect both sit near mirror, and they come from opposite sides of the concept.
Tactics for when nothing works
13. Take a tip at the right moment, not the frustrated one. The best time to spend a tip is when you have a solid green you cannot improve on — you get a second anchor and can triangulate between them. The worst time is when everything is red, because a tip then just replaces the exploration you have not done yet. Set the tip difficulty to Medium and leave it there.
14. Reset your assumptions out loud. If you have fifty guesses and no green, the problem is usually a hidden assumption — that the answer is a noun, that it is concrete, that it is a thing rather than a feeling or an event. Deliberately guess one word that violates each assumption: an emotion, a verb, an abstract quality, a time word. One of them frequently comes back green immediately.
There is no shame in the give-up button either. Revealing the word and then reading the ten closest words on the answer page is genuinely instructive — you get to see the neighbourhood you were failing to find, and you will recognise its shape next time.
What is not worth doing
Four habits that feel productive and are not.
Counting letters. There is no letter information anywhere in this game. The answer’s length, its first letter and its shape are all invisible to the ranking, and no amount of staring at the board will reveal them. If you catch yourself reasoning about spelling, you have imported a Wordle habit that costs guesses here.
Reusing yesterday’s answer. Consecutive answers are chosen independently, so yesterday’s word tells you nothing about today’s. It is a comforting opener and a wasted one.
Guessing the same word in two languages. Each language runs a separate model and a separate word list, so a word that scored 40 in the Portuguese edition has no relationship to its English cousin’s rank. Play them as separate games, because they are.
Working a category alphabetically. Once you have found a warm category, the temptation is to enumerate it — every tree, every colour, every emotion. Enumeration is slow and mostly redundant, because adjacent members of a category cluster tightly and return nearly identical ranks. Two well-chosen probes from opposite edges of a category tell you more than fifteen from the middle of it.
Habits worth building
Keep a mental note of which categories the game favours. Across a month you will notice that concrete nouns with rich physical associations — anchor, compass, kettle — are common and gentle, while abstract social nouns — rumour, nostalgia, sobremesa in the Spanish edition — are rarer and brutal. Recognising early which kind of day you are on changes how much exploration is worth doing.
Play a few archive puzzles when you have time. Practising against a known answer lets you check your reasoning immediately: guess, then reveal, then ask why the ranks fell the way they did. It is the fastest way to build the intuition that makes ordinary days quick.
And when you want the tactics assembled into an actual procedure — what to do first, second and third, and when to abandon a lead — the strategy guide sets out the full method. Otherwise, today’s puzzle is waiting, and there is only one way to get better at it.